期刊
COMPUTER NETWORKS
卷 148, 期 -, 页码 231-240出版社
ELSEVIER
DOI: 10.1016/j.comnet.2018.11.031
关键词
Time series classification; Shapelet transform; Shapelet selection
类别
资金
- National Natural Science Foundation of China [61872222, 91546203]
- National Key Research and Development Program of China [2017YFA0700601]
- Key Research and Development Program of Shandong Province [2017CXGC0605, 2017CXGC0604, 2018GGX101019]
- Young Scholars Program of Shandong University
Time series classification has attracted significant interest over the past decade. One of the promising approaches is shapelet based algorithms, which are interpretable, more accurate and faster than most classifiers. However, the training time of shapelet based algorithms is high, even though it is computed offline. To overcome this problem, in this paper, we propose a fast shapelet selection algorithm (FSS), which sharply reduces the time consumption of shapelet selection. In our algorithm, we first sample some time series from a training dataset with the help of a subclass splitting method. Then FSS identifies the local farthest deviation points (LFDPs) for the sampled time series and selects the subsequences between two nonadjacent LFDPs as shapelet candidates. Using these two steps, the number of shapelet candidates is greatly reduced, which leads to an obvious reduction in time complexity. Unlike other methods that accelerate the shapelet selection process at the expense of a reduction in accuracy, the experimental results demonstrate that FSS is thousands of times faster than the original shapelet transformation method, with no reduction in accuracy. Our results also demonstrate that our method is the fastest among shapelet based methods that have the leading level of accuracy. (C) 2018 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据